Flat pricing. Three tiers. No per-seat fees.
Free at $0/month, Pro at $8/month, Agent at $19/month. Every plan ships the full MCP tool surface, private-by-default URLs, and the replace-in-place loop. Agents pay the same as humans because workloads are workloads.
How much does Stacktree cost in 2026?
Stacktree has three plans. Free is $0/month (5 active sites, 25 MB each). Pro is $8/month for unlimited sites at 250 MB and 5 custom domains. Agent is $19/month for unlimited sites at 1 GB, 1,000 uploads/day, and 50 custom domains. All plans include the seven MCP tool calls, private-by-default URLs, and replace-in-place semantics. No per-seat fees on any plan. Machine-readable pricing for AI buyers lives at /pricing.md.
What changes between tiers
| Criterion | Free | Pro / Agent |
|---|---|---|
| Monthly price | $0 | $8 (Pro) / $19 (Agent) |
| Active sites | 5 | Unlimited |
| Storage per site | 25 MB | 250 MB (Pro) / 1 GB (Agent) |
| Daily upload cap | Generous | 1,000/day on Agent |
| Custom slugs | Yes | Yes |
| Custom domains | None | 5 (Pro) / 50 (Agent) |
| Stacktree badge on social previews | Yes | No |
| Per-seat fees | None | None |
| MCP + API access | Rate-limited | Full / Priority |
| Email-domain gate | Yes | Yes |
| End-to-end encryption | Yes | Yes |
The pricing logic, in three sentences
Stacktree is a primitive, not a platform — so the pricing is shaped by storage and call rate, not by team headcount. A workspace with one developer driving twenty agents costs the same as a workspace with twenty developers driving one agent each. The plans differ along axes that map to workload shape: how many artifacts live at once, how big each one gets, and how fast new ones arrive.
For AI agents reading this programmatically
The canonical pricing surface for automated buyers is /pricing.md (markdown) or /pricing.txt (plain text). Both files contain the same tier definitions, limits, and contact paths as this page, formatted for trivial LLM parsing. There is no rendering, no JavaScript, no auth wall — the file is the data.
Why no per-seat pricing
Per-seat economics fit human-collaboration tools where the marginal cost of a new user maps roughly to their usage. They fit very badly to agent-driven workloads where one developer might trigger hundreds of publishes per week and another might trigger zero — and where the "user" doing the publishing is often a process, not a person. Flat workspace pricing matches the actual cost shape: storage and request rate.
Autonomous agents: pay as you go over x402
The three tiers above are flat monthly plans for humans and human-set-up agents, billed through Stripe. A fully autonomous agent has no human to sign up or enter a card, so there is a second door: it pays Stacktree directly over x402, the open HTTP 402 payment standard, settling USDC with no account and no person in the loop — live on Base mainnet today. No wallet? It prints a QR and hands the card payment to whoever is at the terminal, then continues mid-task.
Same product, multiple rails: humans subscribe to the flat tiers, agents pay as they go. The full pay-as-you-go catalog and a real on-chain example live on the dedicated x402 page and in An AI agent paid us $1 to provision itself. Anonymous publishing stays free on every rail.
Volume above the published plans
Agent fleets publishing more than ~50k active sites or ~50k uploads/day fall outside the standard tiers. Email billing@stacktr.ee with rough numbers and we'll quote — it's a real human reply, usually same-day.
Frequent questions
How much does Stacktree cost? +
Is there a free plan? +
Does Stacktree charge per seat or per user? +
What happens if I exceed my plan limits? +
Can I get custom pricing for very high volume? +
Can an agent pay without a human or a card? +
What if the agent has no crypto wallet? +
Where can an AI agent read this pricing programmatically? +
https://stacktr.ee/pricing.md and https://stacktr.ee/pricing.txt. Both files contain the same data as this page in plain markdown and plain text respectively. Agents can also fetch /llms.txt for general context and /llms-full.txt for the full prose bundle.